1:01:04 Webinar
Everpure Fusion: Scalable Storage Management Without the Tax
Join us to see how teams can provision up to 90% faster and eliminate array-by-array misconfiguration.
This webinar first aired on November 5, 2026
The first 5 minute(s) of our recorded Webinars are open; however, if you are enjoying them, we’ll ask for a little information to finish watching.
Click to View Transcript
Hello, everybody, and good morning and welcome to another Everpure TechTalk. Today's session, we're talking Everpure Fusion: Scalable Storage Management Without the Tax. But before we jump into that, just a couple house cleaning items, or housekeeping items, maybe some cleaning, who knows?
If you've got any questions- Got time you know, please throw them in the chat or the Q&A window. Preferably in the Q&A, but, you know, either one is fine. We will be monitoring that real time, and we'll do our best to try to answer questions live if possible, because sometimes Zane and I get a little verbose and we might not be clearly saying
what we're- Ever what we're trying to say. Ever. But with that, let's actually introduce today's presenter, and we've got Mr. Zane Allen. Hel- hello, Zane. Good morning. Hello. Yeah, it's not, it's not that we're verbose.
Just enjoy having conversations with you, Jason. Yeah. Yeah. You just- We meander it's always a great time. We meander. We meander. And hopefully, hopefully it's useful to the audience. We meander. But, but yeah, I'm, Zane Allen,
a principal technology strategist with Everpure, and glad to be joining you all today and talking about one of my I should say it's my favorite topic, but it's definitely something I talk about day in, day out, so, we're even gonna try a little live demo at the end, so that's always fun. Looking forward to it. Yeah. So I think set the table
for the audience, you know, Zane's gonna kinda run us through the Everpure Fusion, kinda where it is, what you can do with it, but then, you know, besides slides, he's actually going to supposedly he's gonna actually show it to us, right? So it won't yeah, it won't, it won't all just be slide where. But Zane, you know, these calls, we have a mixture of potentially existing Pure customers
or people that are looking into Pure. Like, I think we need to set the table a little, just a little bit, not too much. Yeah. But w- 'cause you're gonna say stuff like EDC and the Everpure platform, and you're gonna say Fusion and these other words on the slide. So give us a, just a, a level set.
Like, just, you know. Yeah. Per- for- perfect. Yeah. You, you, you mentioned EDC. If you've been around Pure in the last, definitely last year, we have something, a concept we call the Enterprise Data Cloud, and really what that is is,
is a model that we are kind of like guiding our customers through how to build a modern data center to really deal with, you know, all these changes, changes in software models, changes in infrastructure, and really boiling it down to, it's the data that's the most important part about your infrastructure. So, you know, all the investment you do in your infrastructure,
especially in these times, needs to lead to, you know, those kind of outcomes, right? It's going to change. Who knows what it's gonna look like in the next year? So I gotta be able to support that, sometimes even on demand, with my infrastructure. So again, it's all about the data.
And what we have are these, kinda these layers. I kinda go above this, but since we're gonna be talking about the one that's highlighted here, the intelligent control plane, really where this feature we call Fusion lives, but it is built on those layers below. So if you are used or familiar with Everpure, you know, formerly Pure Storage,
moved to Everpure this year because of the fact that we're- Yeah concentrating on the data, it's really built on that Evergreen Architecture. Right? Our modular design, our, our, you know, not using SSDs, our DirectFlash modules. You know, the storage gets better over time.
Never have to lift and shift. A united, a unified data plane that sits on top of that, meaning, and, I've heard it explained this way and I love it. We're basically a key value store below it, right? Hashing out all the data.
And then our, our, our, queries are basically block, file, and object. So a, a unified data plane, that means I don't have to buy different storage types for different use cases. I really have those protocols across either a scale up or a scale out architecture. Yeah. Basically,
it's the same operating environment. Like you- Same operating environment. Yeah you know one, you know, basically know the other one. Yeah. And so when you lay this intelligent control plane on that, to that point, I'm now able to address it and talk to it
as a single plane across the fleet, no matter what my fleet looks like from a, from a pure, you know, product perspective. And technically, gonna talk about MCP later, you can even bring in other storage, right? We're not gonna be so arrogant to think that some environments may have a mix, at least for some time, but really kind of bringing all those parts together,
but hopefully highlighting how we are easier to address and, you know, much more scalable, going into the future. Sweet. Yeah, that makes sense. Does that cover it? Yeah. Yeah. I think, I think that's a good I mean, we don't wanna spend too much time on it,
so I think that's perfect. Yeah. Yeah. So. Yeah. And so Pure1 and Fusion are there. Pure1's our, kind of our cloud product. I will kind of highlight how those both go together. Fusion being this feature that is built right into the array,
and first and foremost lets you see a whole fleet at, at a, at a, you know, in one, one connection. Kinda like this maybe. Kinda like that. So- Nice. Look at that. Yeah. Pick- picking up on the cues. Yeah. So essentially, you know, we talk about how do we scale this?
This, this works with one array, two arrays, three, but you know, up to hundreds, hundreds of arrays, right? At different sites and things. So as that environment scales, we don't have to change the way you address it or manage it or get that information out of it.
So it's as built-in, I can't stress that enough. We're not a- making you install some other type of, cloud control plane- Yeah even though it does tie into our cloud control. This is built into every array. I mean, quite simply- So there's not like a separate VM or there's like something
like- Exactly. There's not like this- Yeah control. Yeah. I might, I'll maybe not use the word control. There's not a brain box somewhere. Yeah. Yeah, that you have to maintain and- Yeah give compute to. We try to, in, in our fashion,
simplify everything, so why don't we just run this on box and it simply, it simply allow you to have a picture of your entire fleet. And you can customize these Hey, I want these to be in a fleet. I'll talk about topology groups in a minute here, too. Like, you can organize them, but again, it's just allowing the different arrays,
FlashBlade, FlashArray, right, scale up, scale out, to, to address each other, talk about each other, and then let me see objects and manage remotely across those, no matter which one I remote into, right? And, and, and depending on the, you know, the level of, of, of control that I have, right, from my, from my, profile.
Yeah. Yes. And I guess for me it's a key thing to call out, and it's not a license key you have to install or anything. Like, it's- Good point. Yeah Zane's got a couple of shows just, like, how to set it up, but it's not like he said it's included, so it's not like we're charging extra for it
or whatever. Like, this is all in a FlashArray or in a FlashBlade. There's no license key to, to check or anything like that. So when we say it's included, it's included. Yeah. I used I even like used to say I don't like
to call it Fusion 'cause it's just Purity. Um- Right I'm gonna come around to that. I think we have to call out, like, what the feature is, but essentially as we build more and more features into this, because we, we are basically a software-defined storage that we just put on our purpose-built
hardware, or the big iPhone of storage, I suppose. As those features come out, and we release features monthly, and we roll them up into a long life release every eight to nine months, those feature sets just become available to the customer- Yeah to the end user. So we're building on this all the time.
But at its core, it's a, it's a fleet-wide view of, of your storage. Yeah. I can do this now, right? There we go. You have the power. I have the power. So talk about fleet management, right? Hey, if I can address one storage array, I don't really have to think about storage
FlashArray 1 in 24. I can say, here's my fleet. What's in here? Kind of the next part that it solved immediately was, was this issue, right? This manual placement and, you know, kind of the risk related to, you know,
having to set up everything every time I've I provision something. So, you know, is it the right location? Do I have the right data protection behind it, availability? Is the performance correct? You know, I even like to say from our perspective,
you can set performance down even though it's running on a higher array so that you can give basically different workloads, which is kind of the construct that's coming here. Different performance profiles based on what they need, so they only take the resources they, they require, even though they may be running on one of your higher-end storage.
By that I mean, you know, you have an X, we have an X line and a C line. Maybe you want it to always run at a C performance, but it's sitting on an X for some reason. And then if I did move it over to the C, you know, I keep that consistent performance and there's no, you know, no It's not It's tr- transparent to the end user.
Yeah. But really, how do, how do we address that, right? Is how do we place, take that, that manual piece away, that run, that run book, right? So that every time I provision something, I can decide upfront what that policy looks like. I know that it stamps out the same every time.
And that is something we built into, f- Fusion as well. And we'll talk about in a second, we call them presets, and then I'll give you an idea of where we're going with it, 'cause we're always, always improving. But it is You will hear sometimes it talk about being automation.
I'd like to kind of address this quick, because automation usually is something in the scripting realm, right? You got Ansible, I do something like that, and it isn't that we're replacing that. I'd say we're more augmenting this, right? So automation, usually it's scripts.
You have, hey, what's automated? These are the steps that I always take. Maybe I have some variables, but I just wanna hit go, and then those steps play out- Yeah right? Still building on it. But there's still logic to it.
There's the standard, right? Those are those things that the customer or the end user still has to do. And, you know, if things drift, you know, you have to kind of notice that and then run that script maybe again. Whereas when you go to a policy-based, which is where we're going, so again, we'll talk about presets and then policies,
that logic you can start offloading onto the storage array, right? We're the stewards of the data. We know where it is. We know what your fleet looks like, right? We just mentioned that. I can now Or s- Pure, Everpure can now give the end user or the customer, the ability to go, "Hey,
I want to push this policy out. I predefined it. You know, this is what it needs to be, this is what it needs to run." And now I can get insights from my storage to say, "Great, this is probably the best place to run that. Hey, you have space over here.
Why don't you run it on this array?" Et cetera, right? Get that logic so that the automation, the scripting, or even the manual process, you don't have to go out and do that. Yeah. I don't have to look at spreadsheets. I don't have to go look at dashboards.
I can just kind of ask the storage to say, "Hey, you know what I want," 'cause I set it. Where should I run this? And then drift can start to be visible because I can see workloads in there and how they're set up, right? If something changes, I know that it's from not from the standard, et cetera, right? So just again, giving more allowing the storage to do what the storage can do for you
to kind of automate that. So a- augmenting automation or helping- Sure on the manual. Yeah. And stor- and drift is I, I mean, I know you had it on the previous slide. I mean, we- Yeah you talked about placement and performance, but, like, configuration drift in, especially at scale, is a real problem.
Yeah. Right? Like, this array, a- array one needs to look like array 100. And if I'm doing all that manually, like, well, I'll speak for myself- I'll get- I guarantee you I'm messing up at least one of those. Yeah. We're human. Yeah, yeah. I'm not, you know,
it's not- You know? We know that you're passing judgment on you. It's like, this is just how it is. Yeah. It's hard. The odds are good that I'm going to mess up one or two of those- Yeah you know, 'cause I'm just gonna fat finger. So, getting in there and making sure the data, you know, the, the,
the scripts and the policies and these other things are consistent across your, your fleet, and I know you keep saying that word, but let's say your, your organization of arrays- Organization and that's- Yeah, that's, that's a good way to say it. Yeah. Yeah. Yeah. That's, that's a great And then so
you're maybe spending more time auditing and checking, and, and, and, and, than you need to if you could just make sure that it's easy to see up front instead of front. Yeah. Well, and also with troubleshooting, right? Like, it's one of those things where- If you know everything looks the same,
it re- if you have a problem, it helps- Yeah in theory, in troubleshooting, 'cause you're like, well, you've re- you've reduced the number of errors, 'cause all these should look this way or whatever. But anyways. And, and if policies change, how do I go through and make sure everybody's in the same bucket, right? Yeah. Right.
Which- Right you know, they never, nothing ever changes. Yeah. And then of course, this, this can lead to at a full autonomy, and I'll talk about MCP at the end, but, maybe an agentic future, if you wanna call it that, where maybe if the system knows the policy and I can understand the drift, I can get recommendations from re- remediation,
and then I can ask the storage to take action. Or where we're going with this is I have added a new array, or I've just kind of front-loaded too much on one storage array. Maybe I can rebalance and move that over without disrupting my- environment, right? Make sure everything's connected.
I can let the storage kinda tell me, "Hey, this would be a better place to run this. Has the same host connectivity, et cetera. I'm gonna go ahead and move that for you if you allow me." Yeah, so, so okay. Again, kind of managing I mean, it's almost a cloud-like model, I guess is one way to say it.
You know, I don't The infrastructure informs me what I can do. It understands my intent, and I can, I can take action. I think intent is, like, my new That's a good, that's a good word, right? That my intention is to do this. Yeah. So, just kinda quickly,
this is just a graphic representation, but today how we do this are with presets, and what presets are exactly that. It's a collection of, hey, this is the, the name of it, right? I'm gonna deploy an Oracle, SQL, whatever it is, and, a VMFS sh- data store. Every time I do so this works for file too, I should mention.
This is actually probably very useful, 'cause we run file protocols. Where I could even say, hey, this file type is on a scale up, this is on a scale out, and when I push the preset, it knows where to go. So again, I only have to decide that, right? It's already predetermined, the intention.
But essentially it's a collection of, of those policies. Hey, I have this QOS, I have this many volumes, or I'll let them pick. It's always on this storage class, et cetera. It's got this data protection. And then I can stamp that into, you know, think of it like a catalog.
By the way, I'm showing a GUI on the side. There is the CLI below. Everything in Fusion is addressable by API, CLI, and, and UX. So there, there is no, like, separate control plane, right? We've already said that. It's still Purity on the b- on the bottom end.
It's just giving you that one control plane to, to talk to- Oh, you stole my question. I was gonna be like- Oh, damn. Yeah. But it's good to know for folks who want to script or whatever, you've got the CLI, and for people like me that just wanna click the mouse, that you've got the UI representation as well.
Yeah. And go to UI, and actually you can export these presets as JSON and upload them into other fleets, so it's, it's very flexible. And everything that Purity provides underneath is, is there. But essentially you, you kind of deploy these, make these presets,
they go into a catalog if you will, and then what you do is you deploy something called a workload. So when I use a preset, I get a workload, and what's really nice about that is it's a collection of all the objects. So prior, maybe I'm doing a volume. I have to go find that volume. Maybe I gotta label it, right?
I have a tag, and that's fine. But now I can tag a workload and say, oh, all these volumes relate to that workload. So when I go into the workload, I can drill right down into those exact resources, and just get, kind of get to things faster. Or I can tag them. Maybe there's a department or a, a project,
so I know what, what the, I dare say chargeback. I think that's a very hard thing to do. But at least showback, right? Or- Yeah scareback, I've heard that. Or shameback.
You know, whatever- Shameback whatever you use in your organization. Hey, you're taking these resources. This is what it costs to run this. And I can see that 'cause I'm in a collection. I think I can count on the n- on one hand the number of companies that I've seen actually do
a chargeback and then actually get- Yeah, yeah like the actual charge from like another department. Great idea. Yeah. I worked at a really big organization- Sound- sounded great. Yeah. Just had a budget every year. I'm like, yeah. Yeah.
Well, I used to say that, you know, running an IT is like, you know, I don't need any of this. This is all for you. Right, right. I don't cost you money. I need a laptop. That's, that's all I need. I provide you a service. Yeah.
But anyways. And this little fun little, diagram of how that kind of works. So already talked about it. Got the presets. I want an outcome. Here's my intention. Fusion, Fusion bot.
It's really running on every array. We'll say, great, I understand that need. Maybe I even went to Pure1 to get that placement recommendation. You take it, and then it forwards the commands onto the correct array. If you're not logged into the right array, it'll just forward that on.
So I kind of go back to the file example. I think that's really powerful is, hey, these files are always scale up, and these are always scale up. But the, the person's provisioning just has to choose, like, I need this file type. It's a, it's a user share.
And it goes, oh, I know what storage class that should be in. I'm gonna put it over here and provide it, provide you a, an export, you know? So bang, bang, bang, and then everything on the array just is now managed in workload. So again, I can drill right in, I can see what, what makes them up. Very, very, very easy to use.
Built right in. So we're always innovating. So this is kind of a larger diagram, but I'll kind of explain this. Presets today, presets go to workloads. Somewhere where we're moving to, to make this even more, we think kind of to fit that,
you know, the, that management, paradigm, is what if we were to break this out into more of a policy-based? Meaning, I know or maybe there's somebody in my organization that knows what those individual policies should be, right? Here's IOPS. We have these different types of workloads,
bandwidth. This is the latency profile. More importantly, probably recovery point objective, recovery time objectives. Like, you know, what does the resiliency need to look like? I need to burn, burn, build those service levels in because maybe that's how I have to present it to my end user, my customer, right?
This is what I provide for these types of workloads. And then of course, other config policies, naming, you know, snapshots, all these sort of things. And you kind of build all these standards, so you can maintain a set of standards, and then they make up what's called a workload definition. So think of the preset I just talked
about turning into workload definitions. So now when I have a VMware, an Oracle or a test dev stuff, as I see here, I can pull from these SLAs, these config policies, build that workflow definition, and then again stamp out a workload from it. So the catalog sits in the middle, it's made up of all these pieces,
and then it opens up a, a large world where we can say, "Hey, this SLA changed because of some reason," or, "This new SLA was introduced, and now I can see the drift from a workload definition, and then maybe what was pressed from it, and I can correct that and update. Or I can set up a new policy," right?
Just an easier way to maintain than having a new set of presets, right? Again, setup and, and definition intention is coming from the SLAs and config policies. Workload definitions are stamping out what we need for work, and then the workloads are what we're actually producing from that catalog. So again, s- same interface, just kind of an update you'll see, here not in the too,
not too distant future to bring those presets into an SLA or a, a, a policy, an SLA policy construct. Completely able to upgrade from one or the other, if that's a question. So we won't, we won't make it, "Hey, you gotta reset over here," right? We'll be able- to bring it in here.
Just wanted- Redo everything to throw that out there. Yeah. Don't wait. You don't have to wait. Like, get started- Yeah with presets today. Get this tomorrow. Yeah. All right. So just so I, I wanna make sure I g- I,
I captured that- Yeah saying Sorry. So on the far left, the define the standard, like that's me at the keyboard or the CLI saying, "Okay, I'm gonna build a service level." Like you said, you've got, I want performance. I'm setting up availability, so I'm saying, five-minute RPO,
you know- Yeah one-hour RTO. Like, I'm setting these in the, the S- SLA. Yeah. But then I'm going into something, into a To create the workload definition, that can be- Think of it as the preset. Like the upgraded preset. Yep.
Yeah, and that can be a grouping of multiple service levels maybe or multiple config policies to kind of bas- basically, it's a, it's a bucket that's holding a bunch of the s- standards, for lack of a better word. Exactly, yeah. And- Okay yeah. Yeah. And then once I've created the catalog, now I'm just saying,
okay, now when I'm provisioning these new volumes, that would be in the workloads and they would follow the catalog? I just wanna make sure I'm following, like how I would, you know, if I was sitting down- Yeah, correct. Correct, correct, correct like I was deploying this. You would, you would Think about it,
and you It could be the same person, but we'll actually have a new user account as well, or a new, ser- policy. Sorry. Yeah, a new user account that'll, could just define service levels. Mm-hmm. So, hey, these are our standards,
and I define them, and then there's maybe a group, and it could be the same person that says, "Okay, great. Based on those policies that I'm provided from the business, this is how I'm gonna build these different workloads." Got it. Okay. So yeah, pull those pieces over. It's a pick list. And then the workload definition
is the catalog that I Hey, I push- Got it I pick it, I get the workload. Exactly. But it Then it means that you're only managing changeover in a SLA and a config policies perspective. That can update and, and inform the catalog. Of course, you could make new things in the catalog too, or you could just say, "Oh, okay,
this SLA changed. Our Oracle server needs a better resiliency strategy or RTO." Yeah. Or think, you know, we have a lot of, data protection built in. Like we just released our ActiveCluster with Active DR, right? So now you have your HA, no, no zero data loss, complete flip over to zer- near-zero RPO
for recovery, right? That's something that just came in this year that's now available. I can now go update the service level and, hey, I definitely wanted that for my Oracle, so let's go ahead and update that, right? So now it's just a lot easier to kind of flow in changes.
Got it. Yeah. So you're also getting a many-to-one on the standard to the catalog. Yep. Yeah. Got it. Just a lot easier way to kind of manage that. So just to kind of hit that two, two requests, the reason why,
kind of what we're addressing here is, you know, steps by hand, all of the kind of process. Of course, the drift on the end I think could be the most I gotta keep up with it, to bringing policy where I can request from the catalog. I know it's up to the standard, or I can at least check it against the standard.
It's protected 'cause, you know, everything's on there at creation. And you can monitor that drift. So automation from a sense of what storage can provide from an automation. I know what you intended. I know it's landed on the storage. I know it's running in the right optimal place, and I can inform that backup,
and then everything above that can be either manual configurations or usual Ansible scripts or whatever the organization's already using. Few things. This is actually in the product today. It's gonna get better as well, is something called topology groups. So we, we mentioned fleets.
We just released, just give the number, 128 can be in a fleet today. That number will keep ticking up. It was just 64. But you don't have to separate things by fleets, you know, depending on the size of it obviously. But you can kind of think of a fleet as my control plane.
I wanna see everything, but now I can also logically group things into different, topology groups. You think of these as availability zones. We just didn't name it that, just to kind of leave it open. You can see that represented here. I got, you know, the regions, and then I have maybe, you know,
these data centers that are inside those regions. I could see this also, I think of like healthcare where maybe my Epic environment, I wanna see my whole fleet, but my Epic environment always runs on these two arrays, and I just- Yeah I just need to know that. My VMs are here, my test and dev's over here, right?
You can see it logically grouped and then, you know, you can, you can take action. It makes, helps you better inform, right? Especially from those policy perspectives, like where to deploy things. Yeah, I think the key Yeah, I guess what I'm taking away, the key for here is- There's probably no one right way to set this up.
It's like how you as the user want to logically think about- Yeah your systems. Like, obviously we're using a very regional-based example here, but to your point, like maybe it's You could do departments, or you could do applications. I mean, it just- Yeah depends on how granular or where you wanna go. But the f- the functionality is there.
This is just an example of, like, how you might break it up from like a If you're a global- Yeah organization, let's say. Kind of the easiest to understand, you know- Yeah. Yeah especially if you're doing cloud, like, oh, I have availability zones. And then you can think of things in, okay, I have re- redundancy or resiliency across.
I understand that because I know this is here, and this is in this region, and that's the distance I need and et cetera. Yeah. Yeah. So yeah, it's just an easier way to look at it so you don't have to go 'Cause if you just had to divide everything into fleets, that'd still be great, but now I gotta think, okay,
what does fleet two do, and what does fleet three do? Like, how about I just have that all in one dashboard? Yeah. Curious, Zane, and, is there a graphical representation of this in the interface? Like when I build this out, like if Do I go in?
I don't know if I'm gonna look for a good or bad answer here, but like Yeah. Well, it doesn't show it like this yet. We are pushing that into Pure1. Okay. So you can see that from a Pure1 perspective. Exactly. In the UI, it shows you Yeah, to a degree.
Under fleets, it shows you- Yeah kind of what's grouped in topology group. You can see- But in Pure1 I could go in and kind of like I'd see- Yeah like a, a map and I could click- Yeah. Map, lack of a better word, but I can kind of drill down through it. So there would be a- Yeah visual representation of how I've logically kind of-
Yeah logically set this up. Yeah, yeah. Yeah, it's not just all behind the scenes and I'll figure it out. Yeah. Yeah. No, that's, that's a good point. Yeah. And that, that's gonna lend itself to, like, where we're going with it.
So think about it just from a policy perspective. If I can seg- or I can restrict policies to different topology groups, that makes that a lot more useful. So I have policies, they're set. Maybe I can, you know, ring those and say, "Hey, they can only be applied
in this topology group. This topology group has these policies," right? So you can start enforcing things. Again, this is when we're talking about scaling, is I have a fleet, it has topo- topo groups. These placement policies are at the top, and now I know where I can and can't deploy
things. Or from a resiliency perspective, I need to have this, you know, recovery point objective, but I need it, you know, in ADR. It needs to be out of the region, so that has to be these two AZs, right? Or these two topology groups. I can never replicate inside of it or replicate over there, right?
Just making sure I do the right config up front, and I don't have a lot of manual thinking about it, and I don't skip something or, you know, get into trouble later. So you can set policies and then set the boundaries of where those policies should be placed, right? Yeah. So again, automating storage to give you that info.
Like, hey, you want this? Well, this is the pair that you need to replicate to. Hey, you want this intent? Oh, that can only be in this topology group, 'cause that's where the XLs are, right? What- whatever the case may be. Or the, or the healthcare example, right?
So you can start seeing how these can help even more just graphically. It's like, oh, I can start enforcing based on these topology groups. Can I set security boundaries with this too, by chance, Zane? Like, maybe I've got a group of admins and You know, well, let's go back to the logical layout, 'cause that'll make it easier.
Maybe I got a group of admins in the US, and I've got a group of admins in the MIA, but I don't want them- Not- To ma- Like, could I But I'm like- Yeah, I know what you're saying. Like, can I do RBAC? I'm the grand poobah sitting over the top of it. Can I- Can I do RBAC by, by- Yeah topology group?
That, that's more a realms construct today. Okay. All right. But, but- So, so- But that, that isn't out of the realm- Possible, but just not- of where we'll go. Yeah, yeah in a different area. Yeah. Yeah, yeah. Okay. Yeah. That, that will be interesting to see
how these, how these group together. But yeah, that, that'd be more of our realms construct. Sure. Okay. Yeah. Which- I could just see kind of that crossover a little bit. 100%, yeah. Like a fleet-wide realm- Yeah that does by topology groups. Yeah.
Yeah. Hey, this realm Yeah, that would be perfect, especially if you're, like, replicating. I'm doing it into this realm, 'cause this is all I have access to, and here's the three topology groups that I can go into. Yeah. Yeah, yeah. By the way,
do you get the grand poobah reference? Is that- Yeah. Is that leading me up to- Jason, Jason knows I love it Royal Or- Royal Order of the Water Buffalos. I got my hat over here. That would be cool. Just put that on.
Be like, "Whoa!" Anyway, all right, we digress. So again, I'm, I'm kind of mixing where we're going with this. These things aren't too far off. But now with these policies and workloads, and I'm deploying things predictively, or my intention, right?
Hey, this is what these policies need to be. It makes it a lot better to kind of predict and look ahead, right? Yeah. And that's also where we're gonna get informed for moving storage around is, hey, you're gonna run out of room here, the headroom for your performance, whatever it is. We might wanna migrate, which is relocate at the end.
Something we've been hinting towards, coming, a too distant future, where I can say, "Hey, this is kind of a hot zone. I need to move things over." And if I have those things in workloads, I have that intention, I know defined parameters of what's good looks like, then the storage can predict, hey, I can move this over here and still maintain your SLAs,
right? So you gotta kind of get that construct in place first for a relocation to make sense, because it's really like, hey, I want to relocate this, or maybe you- you're saying I need to. If I run it over here, do I still get the same performance SLAs that I need? Yes. Okay, take it. Ne- never should it be automatic,
which I appreciate. You know, you definitely wanna know what happens when I hit the big re- bu- big start button, but- Right let me make the decision. But you'll, you know, let the storage inform you and give you all the checks that it can, it can fulfill what you intend, if it moves over.
And then you can kind of, you know, and then ob- obviously optimize day zero. We talked about that already. I can I know your intent, so when you ask for it, I can give you all the places that this would run, exactly as expected. So you're not, "Hey, let's put it over here," and then now I have a, you know,
something got slow or availability isn't there, now I gotta go troubleshoot it. Let's just make sure out the gate it's doing exactly what you intend it to. So long, gone are the days of just make a volume and attach it to a host and all that. So- Yeah guess before we move on, I don't know if there's questions out there or anything, or anything else to set on 'cause we're kinda transitioning here.
Hopefully that makes sense. There was one question that's kind of related to this, so I was gonna let you- Okay go here, and then I'll ask the question that came up. So So we talk about Fusion. We talk to a lot of customers about getting into a fleet, and it's always like, "Hey,
what do, what do I get out of it?" And kind of just spend some time walking through what, what it can provide. I think fleet-wide view. By the way, if you're still doing, like, I just wanna provision volumes by volumes, maybe I'm not ready for presets. If, if you go in to make a volume, it's still Pure Vol Create.
What's kinda nice is in a fleet if you have things like host groups or even pods, if you're using our pods for, like, replication, you can say, "Hey, I wanna deploy by host group," and it'll only give you the arrays that have those host groups connected. So it even provides just some benefici- benefit just there, right?
I see a fleet, and even though I'm provisioning a volume, I have a better view of my entire fleet to do so. So, so still some value. But what I've been talking a lot lately is about the MCP side or, you know, as we put here, point your own AI at it.
I wanna caution, I'm not asking I'm not advocating you just let AI manage your entire storage, though that could be a future, right? You know, hey, AI optimize storage so it deletes everything. But really- Optimize space, delete. Got all of your QOS problems done.
There's no biz- noisy neighbors. You know, you're on- you want the recommendations, you don't want the, hey, take actions. But it's really this idea about the MCP server and really easy to A- again, another thing that's just built in. Talk about the couple that we have, and I'll hopefully show you some, you know, that, that,
that, that demo holds up. Hold Show you some examples on, you know, where you might use it. But essentially it's natural language. Hey, what if you make a fleet, you install this MCP server wherever you like. In fact, my demo, it's sitting on my laptop.
And then you can just ask your storage questions. Hey, I need to know this. This is the data I need to know. I know it in words. Ask it to your storage, right? And get those results, and then kind of dig into it.
So when Maybe, I don't know if w- if you wanna answer it after this, Zane, 'cause it ties into your- No, go ahead what you said before- Yeah in that previous slide. As somebody had said- Yeah a- and I, trust me, I know this is kind of a slippery slope, for lack of a better word, but, you know, I don't love Somebody said, "I don't love the idea of AI deciding where my databases live," right?
Sometimes people, that'd be like AI, you know? Anyways, if Fusion recommends a placement I disagree with, what happens? Oh. So I'm assuming if the AI- Yeah, yeah. Go for it or somebody says, "Hey, you should move the Zane volume to array two," you'd be like, it just doesn't do it, to your point.
Like, it's gonna say yes or no, right? Right. Like, you're gonna be able to agree or disagree. Yeah. It's informing you, and it gives you- There you go the reasons based on your intentions. That's why I like that word intention. Yeah.
I'll say it a lot, is, hey, I know what your intent is for this workload, so I recommend this, and but you, yeah, you have full autonomy. You, you are the auto- you know, it doesn't have- Yeah full autonomy. Yeah. You have agency, I guess would be better to say. Yeah, so even in two-part, when you're provisioning, say,
you can say take recommendation, or you can say, "I specifically want it on this array." Okay. Yeah, that's, that's, that's, that's, yeah, that's, that's never, that's not going away. And where I'm talking about here, actually, you know, there's, there's a couple MCP servers.
They don't You can change policy or presets and such, but as of the version that's out, we don't really take a lot of write actions. It's really more of a, hey, let me, let me view first, right? Mm-hmm. And then I can take action based on what I see, right? I'm being informed. Yeah. Yeah, that's a great point.
Yeah, I- Cautious observer somehow I am Yeah, I wanna, I wanna get to the, to the, you know, the answer fast, and then I wanna take the action. I don't wanna spend the time, like, doing the research and- Yeah you know, 'cause maybe I don't have it, right? Yeah. Yeah, I, I, spend a lot of time in data protection, and remember customers say, "Oh,
so your failover, it'll just do it for me?" And I'm like, would you really want that to happen? You wanna go to a whole data center failover because, I don't know, the network blipped out for two seconds? Or do you wanna know exactly what's gonna happen when you make the move, but have- Yeah,
you want a little bit of control the power to decide it. Yeah. Yeah. And I think that the latter is the better, the better. You wanna know what you should do, how the escalation path works, and that it's a tested, proven method when I hit it, but let's, let's We're still in charge.
We're still in charge. Um- Yeah. And the machines haven't taken over yet. Yeah. Now, I'll kind of talk about getting to agentic. I think there could be a world where you're like, "Hey, here's some things that I just will let you take care of,"
'cause I always do it this way, but it's very- Mm-hmm strict. But it's really a crawl, walk, run situation, I think. So if you're not familiar with, MCP, it's the model context protocol. It's really still using APIs at the back end, the ones that are published. It's been said it's like the USB-C for AI.
I mean, it's just USB-C for A- AI, APIs, meaning put this MCP in there. When I ask it that question in natural language, right, you do tie it into your LLM, and this can be I'll show it in Claude 'cause that's what we use. But, really anything in local models, et cetera, right?
Mileage may vary. It then determines what is the best tool that is available from that MCP, and then takes that action against storage, right? Say, "Hey, I wanna know where these things are." Oh, that tool is this. I'll go out and reach out, get that information, return it to you. We have something called Copilot.
Of course, we've named it that as well in our Pure1 that gives you some of this, too. Like, tell me about my storage from a Pure1 sense. And then, you know, we have an MCP for Fusion, and we have one for, for, um- Pure1 Manage too, so you can do a little, something a little different than Copilot. But basically, yeah, AI picks the tool for you, gives you the result.
And you set this with API tokens. You can have different layer of, of permission or, you know, a role-based access control. You can say the IMCP server can always read and never write, right? There's a lot of controls you can put on this, and it's the same with AI. You get very defined steps.
But it's a really nice way to say, "Hey, I tie this in, and I can just ask my storage questions. Like I've literally just said, "Why storage bad?" Right? Let's just talk in real broken Eng- you know, real bad English and- Yeah and it'll still give you results that it's like, "Hey, I think this is what you wanna know," and then you can drill down into it.
So I talk about crawl, walk, run. I think, and I, and I started thinking about this lately for customers, like, "Hey, I don't need automation, I have everything. Hey, I don't wanna change my process for presets. Fleetwide's nice, but should I take the time?" And again, it's really like, well,
would you just like to ask your storage questions? So low barrier to entry, really. We're not gonna change anything. You just have to have Fleet, and you point this at it, and then you can start seeing your fleet and asking it questions.
Now, the next step could be take action. Like you could say, "Hey, um," show example, data protection will work where I have backups and I have snapshots. There's an API for that, and I have that token set up. I can say, "Great, clone that snapshot for me, and then I'm gonna go take some action on it." It'll reach out and do that if you want,
if you have permission. So there is a place where you might go, "Okay, I'm comfortable with what I'm seeing. I understand that that API or that, that command, or even if I wanna run a script, does this action, 'cause I do that anyway, so I'm ready to just fire that off there in the context of what I'm trying to get done." And it But it does also set the stage
for a full agentic I say full agentic, but a more agentic operational- Mm-hmm model. So maybe there's customers out there that are thinking about, how do I use AI to do operational benefits? Maybe you're tasked with doing a project. Like, this could be a really easy way to get started with that, saying, "Okay."
And I'm working on a model now to kind of build some examples of agents where we break them out, where they have different varying, You know, some are monitoring, some are taking action, and hierarchy. Not quite finished here yet, but, but you might have an agentic environment that says, "You know what? Help me triage root cause of this,
and I'm gonna tie in other pieces of my data center, networking, storage, vCenter," et cetera, and kind of get to the root cause, right? Again, giving you that information so you don't have to take all that time so that you can go take action or open, or maybe even open a ticket for you, right, to get down the line if somebody else handles another part of the business.
So I, I kinda see this journey if one goes on it. Like, yep, I'm just looking. Maybe I'm taking some action, which is real direct. And then maybe I'm building an agentic future. Someone's helping me build that.
And then, you know, I'm approving everything, and then eventually, hey, 20% of this, I don't ever have to approve that, just take action- Yeah open a ticket, do the thing. And I still wanna know about this. So just think of it as another A, it's a coworker, right? A coworker in your environment.
I envision the dog from Up- Yeah. Where you can just be like, now I can go to the array and be like, "Tell me what's wrong." Yeah. You know? And the do- you know, and it can turn, it can talk back, it can translate for me. Sorry, that's kinda how I approach it- That's actually a good point.
Yeah, yeah, yeah. Yeah. I think we talked about this before. Like, if your dog, there's something wrong, can't tell you, right? Yeah. Well, what if it could, right? It has a collar. Yeah. Just tell me what's going on, Bob. Yeah, yeah. Like, I just need to know.
You tell it you're the alpha, though. I'm the alpha, still. Right. Anyways. Yeah. And to kinda land here, I definitely wanna get to the demo part. It is always fun to watch someone try to do a live demo.
But really, I say automation to kind of put it in, but it's automation your way, right? We You can have workflows with Copilot. We actually have a tool called Workflows where you can build steps in and maybe trigger it on results. Copilot, right? We talked about that. There's a,
there's an LLM that's sitting in Pure1 that can give you some actions, right? Hey, tell me my security posture. Getting to that agentic AI future, like I mentioned, or even into this collaboration, AI agent collaboration. Even to just an MCP collaboration, right?
And I'll show that here in a minute, whereas, like, I can tie in vCenter and kinda get a, kinda get an answer across my data center, not just from the storage layer. All right. So I'm gonna launch this here in a minute, but, the two Well, I'm gonna show some others, but two that we have really, are the Fusion and the Pure1 Manage.
There is actually a Copilot one as well, so, like, that, that LLM can tie in. But our Pure1 Manage, as you see, read only. Already roles that you have configured, right? It's basically a, an authentication to Pure1, so whatever you have in that org. If you haven't done that yet, we have an IM for those organizations.
We call them organizations, your little collection. And then of course the Everpure MCP server, which goes directly to the arrays, and is based on a, also an API token that's based on your role, right? In Fusion, we do centralized authentication, and everybody kinda gets mapped to a group. These are available.
So the MCP API for Pure1's from Pure1. There is a GitHub for the MCP server. There's actually a new version almost dropped, I think dropping in about a week, so we're gonna give more tool sets to this. Oh, nice. But, I didn't For the purpose of the,
for the lab, I didn't, Or this demo, I didn't change anything yet. There will be some more write actions in that, so it'll be really interesting. I think a little bit more, operational. So do you want me to stop sharing then so you can- Yeah, but if I could- Oh, sorry this is basically the tools it covers, but let's,
let's just go into it and see what we can do. Yeah. Let's see, let's see this thing in action. So I have-- I'm using Cloud Desktop, like mentioned. Hopefully that's big enough. You tell me if it's not. And- I can see it okay, but if anybody wants to put in the chat if they can or can't,
that would be great Yeah, so this is Cloud Code. It's what I use. So I have this set up. We should be able to go here. And like I said, we can see what I have set up in here. Now, you can do this in desktop.
I've done it in code in the actual install, but I'm doing this in a terminal on Mac 'cause, I don't know, maybe it looks more legit. Again, this can be done with whatever customers are using. I'm talking to more and more customers that I mean, six months ago this probably would've been, like, whatever, but now that we're all kind of running, I think,
some sort of AI on our desktops probably daily, this is why it's just becoming more and more of, of a topic. If I go in here, I can see, see the tools that it gives me, right? Only ones that are destructive, I like that word, are the presets, right? I can go change the Fusion stuff.
You'll notice everything else from a fleet perspective is just read-only. I do actually have Veeam in this one. I have vCenter, right? It has 33 tools, right? So there's just MCPs are kind of available out there, and you can and you can combine them.
But this is what I have set up in my lab. So in the interest of time, I won't make you watch me type. I have set up some, some prompts. So I just wanna see the Fusion fleet, right? What do I have in this fleet? It's gonna call out to those.
You can watch me burn my tokens, and it'll give me a representation of, okay, here's, here's the arrays in your fleet. I got the Purity version. You know, it'll tell me if it can or can't see things, so you can kinda drill into it. This is my one of my fleets that are in here. Next I can a- maybe ask it questions like, okay, well, in this fleet, you know, what,
what are, what are the, you know, what, what are the largest volumes that I'm looking at, right? Maybe I need to know that. Hey, w- who's full? Who isn't full? I just wanna get a, a, a good look at that, right? Maybe your boss comes down and says, "I need to know this thing,"
you can literally just go to MCP and ask that question, right? Well, I don't know, let me copy/paste your email into this and see what I get. Of course, it's gonna take its time. So- One can ask anything here. I can literally say, like, "What do I need to know about my fleet?" It'll go like, "Hey,
here's the errors. You have vast STaaS issues." You know, I see that these ports are down, right? Just anything that it can report back. But this makes nice little Hey, pulled up 10 volumes. This is what they're called. It's funny,
number two and three are called huge, so I guess that makes sense. Here's what you provision. This is what you're physically using. This is a lab, so obviously those numbers aren't impressive, but- You're right but, but, you know, just relatively, what do we have on here, right?
So And it gives you an even interpretation of how I rank these, right? So you're like, oh, that's not what I meant. I wanna see it in this perspective, right? You can really just, again, ask it anything, anything you need. And then, let's say, okay, well, is there any critical alerts across my fleet, right?
I'm just taking real baby steps here, right? You might know the problem you're trying to address and get dir- directly into it, but just- Yeah as an example. Yeah. Plus, if we wrote, like, a very intense prompt, nobody wants to sit here and watch.
Yeah. Yeah, yeah. Watch this wait for, like, five minutes while it's crunching. No, I ran through some of the demos and was like, well, that one will take seven minutes. And I'm like- Yeah yeah, yeah, fuck this. I don't know. Maybe they do. I don't. I mean,
I can enjoy my coffee while it runs, I guess. We could probably just do this. So, hey, show me the critical alerts. Oh, look at that. My gosh, I thought I'd actually see some. Nope, no critical alerts. You're good to go.
That's pretty good for a lab environment. Yeah. I, I did not actually expect that. It's very dynamic. So- So say you, you spend a lot of time talking to customers. Is this When, when customers are using MCP, is this kind of the interaction?
Is it mostly where they're just trying to quickly find information? Is that the most common use case where it's like- Yeah, yeah and again, to your point, they're just using plain words to like- So something slow like, all right, just give me an overview of, like, you know, where, where, where's the most busiest volume, right? What's the most busiest array?
And you can Then you can just start drilling down, so kinda helping you get to, to a troubleshooting point. I can tie this, too, with Pure1 so that I can, And actually, I'll show you a dashboard that I popped out of here as well, where maybe, I think another issue coming up lately is CVEs, right?
Security s- landscape has changed dramatically because of AI, so why not use AI to help you? It's like, hey, what are my Purity levels? What's my CVE? You know, where, where, where do I have those critical updates that I may need? And, you know, where should I update?
Like, what action should I take first, right? So you can ask that question here and get drill down to it. Like, oh, no, we're covered. Oh, it looks like this array's out of band. We should probably plan to upgrade this.
And you can even use our self-service upgrade to do it. This is a place where I've tied into Pure1, so I got the Pure1 MCP. And not only am I asking from a local, hey, what may be critical, but I'm now can getting more of a details of, like, what to do about it, right? 'Cause Pure1 can kinda give you that, that landscape.
So tying that cloud plane that you don't need it for Fusion, but obviously giving you a more informed step. So, hey- Sure this is how you fix this. Your NVRAM device is unhealthy. Here's a support case, right?
It's In fact, they've already opened that support case for you, right? You can just kinda get down to the action easier by tying the two together. And again, they're just both configured. I ask the question, it reaches out, finds the tool for me and relays it. So no- nothing I did here.
You know, I had to It's essentially just, just prompting, right? I do wanna get to the protect one, so let's kinda skip ahead here. And I know we have other dashboards, but Oh, wow, that's bad. Whoa. Sure what we did there I spelled it right That is a color scheme for sure, Zane Yeah, maybe we don't want that one.
Wow. Of course, live demos, that's not exactly what I expected. All right, let's get back. That's better. So boom, right back at it. You can even build kinda dashboards and where these could be great is like, hey,
I wanna see things in a certain landscape. Maybe it's just really a report that I just wanna produce to somebody. I can do that artifact and pop it out, and like every good cooking show, you wanna have one ready. Here's one that I built exactly with prompts like that. I even told it, "Hey,
here's my Everpure branding, so put the logo in there, right? So it looks good. Oh, wow. This one actually goes out and, and, and updates itself, so I don't even have to go into the prompt to get updates in it, and it gives me things in the view that I, that I think is actionable. So I can now drill into, you know, I want graphic representation.
I can drill into a fleet, and I can go into the array and I can see, you know, the issues. I can see what's tied to it. I can understand, you know, who's connected to these volumes. These are the actual fleets. I actually have three fleets in this lab.
This one represents topology groups, so I know things that are in topo groups. This one doesn't have any, you know, et cetera. Here's my open alerts that I can look at. And what I was kind of trying to show with that Pure1 upgrade is I, I asked for a dashboard like, well, show it to me in the perspective of, of the update,
any critical issues and CVEs, right, that I might need to do. Right in the middle of a refresh. That was awesome. Oh, that was That's my other dashboard. So anyway, as we were doing this, this is the dashboard that was created by that prompt. That's pretty cool. Yeah, yeah.
And why I like this one is, again, I think CVE exposures are kind of a big thing, is I can now look at what are my critical arrays, what's the severity, what's the fix that's available, where are my fleets in their update cycle, right? Hey, if I just did an update, I'll get those CVE packages and I'll be, you know, in compliance. Break it down by even, you know,
everybody likes to gamify, so we put a little score here, right? So anyway, the sky's the limit. I just kinda put this into the prompt like, "Uh, show it to me in this value, give me the score," right? All of this is done in natural language.
Even has this little prompt here. I like this part though the best. Hey, this is what you should do right now. This is what you can do this quarter, and here's how you can plan going forward. So you can even help it build j- help you build a plan based on what your outcomes need
to be so that you can say, "Hey, this is what we're gonna These are the steps we're gonna take business to get, to get backup in compliancy. Sweet. That's pretty cool, Zane. So again, I can go here in the prompt. It's fun CLI, but I can also go, hey, give me a graphic representation
of exactly this scenario right now, and now I have something to produce an artifact, right? I can pop that right out. I can change it, you know, g- whatever I need to do, and I can produce that. Or it can just help me visualize what I need to get done. So I'll do one more here to kinda show, So let's say like the AppTain- Yeah,
you got about, you got about eight minutes. Just giving you a time check Yeah. The AppTain told me that Test VM 01 is corrupted about an hour ago. What are my options?
Again, hey, I just got this information. Great. What, what can I do about it? Can you make that full screen, Zane, by the way? Oh, yep. Please. Not a problem. So Test VM 1, I'm asking it a VM.
So this MCP is now how we can start branching out here, like how do I tie this into vCenter and Pure? And in this case, it'll actually have Veeam behind it. So my data protection software is I know the result is I need to fix VM 1. Do I have to go take the time, where is VM 1 running?
I could do that, or I could just ask the, the MCP, the little collaboration of them, right? You saw that all those were in connected. And it'll say, "Great. Here's what you have available." You have a FlashArray snapshot, and you can use the Veeam plugin 'cause you have Veeam, or you can clone it, right?
Right from the array and pull out the data you need. Gives you kind of the speed even breakdown as like, well, that'll be minutes 'cause it's just a super simple cloning operation. But maybe you like the time better from restore point for the backup. Great. You can do instant VM recovery 'cause you have Veeam or, you know, if you have Rubrik,
whatever it might be, you can pop that up. And then, again, I talked about the array snapshot, or you can just straight copy it off the Pure array and do it the manual way, right? Here's what's available to you. Here's the last restore point.
You can in a really quick moment understand like, okay, I can take this action, it may take this long, and it'll give me this result, and all I had to do is ask it, "Hey, this VM's broke. What, what's my course of action?" Tie this into local KBs too. Hey, every time this thing pops up, this is what we do about it, right?
Here's the, here's the knowledge base that gives you kind of all the, all the input. This is how you do an instant VM recovery. Which by the way, you can, you can actually, because those APIs exist, ask it to recover from those snapshots. So I can tell the array if I have the rights to just go ahead and make that clone,
or I can, because a, that API exists on the Veeam side, I can tell it to start an instant VM recovery, and it'll kick that instant VM recovery off. So I can kinda go already get to the action and then, you know, go into that console or go into vCenter and, and then, you know, finish the operation. So again, just kind of like getting you right to action.
Hey, I know what the problem is 'cause someone told me it. What are my options? I'm gonna take this option. Go ahead and get me started, and then I'll, I'll roll over and get that done. So, you know, just, just making, like in this case, data protection,
the time to result just that much faster, right? 'Cause that's, that's the crucial part. Like great, I have backups, but if it takes me days to get back from them- Right what's the point? You took the backup, but can you restore the backup?
Yeah. It's all about restore now, right? So I could kick that off here if we had time, but if anybody wants that proven, but I maybe we leave, I don't know if there are any questions or anything else. I just I've got one or two- If people want to- But I just wanna make sure you've got what you've got- Let's go before we- Yeah. No, I wanted to show like,
hey- Before we wrap up let's pull things together, 'cause this is probably where it makes more sense than data center. I need to do a thing. Give me the info, let me take that action, right? More than just like, hey, who's the biggest volume? Even though that, that can be useful, too.
So why don't you go and let me share screen, Zane, and we can kind of bring this home. Yeah. Yeah, fun with MCP. Easy to get started. Nothing's will get broken, but you can see the possibilities it can possibly- Yeah do in your environment. So, we just have a couple minutes, so I'm gonna probably just ask one or two of these
that came in, Zane, and they came in at different times, so we're gonna kind of b- be- do a little context switching. But this came in around the presets. Sounds like they've already, already using some provisioning scripts, so I'm assuming they're doing something CLI or using the SDK or something.
Nice. So what, what do I lose if I throw them out and switch to presets, and can I run both while I'm testing out or looking at presets? Oh. Yeah, yeah. You, you If you have presets, we change nothing about your operations.
What you can do, if I understand what the scenario is right, you can take- It's just about like can I have both going on at the same time. You can have both. You can use your scripts and they can call presets. Dope. So I can say, "Hey, I need to provision this," and it goes down to Fusion and says,
"That's this preset. Give that to me. Now I'll go pull the volume and then pull it into the rest of the operation." Yeah. So there could be some like re-swizzle of the, the, the, the automation to, to, to, accommodate that, but it should make it easier, meaning my automation doesn't have to know where to put it.
It doesn't have to pick. I can take that all out so I can let Fusion inform the automation or the scripting where to, where to provide something. Or- Yeah, so- Or press out a preset and then use automation to put something in it, right? I mean, it's, it's up to you. But yeah, we do not It's not either/or.
Yeah. And we don't change your operations. And it just might be even, you know, how complex those provisioning scripts are, so. Yeah. Like if it has to go out and like- Short answer is- go across all my arrays and- Use both Yeah. But long answer is it depends based off- Yeah, what the- the complexity of what you're trying
to do. Like if you have automation like look at all my arrays and tell me where I have the most space, like eliminate that, right? We can do that with Fusion. Yeah. For instance. Yeah. I think we have time for one more, maybe do this in 45 seconds or less.
Talk about moving workloads. You mentioned like moving workloads. Yeah. Moving, moving a production workload between arrays usually requires a maintenance window for us. What does that look like with Fusion?
So that, that's one of the, the future states. What we'll essentially be using is, think about it, to make it simple, is kind of our ActiveCluster scenario. So the intent is the migration happens non-disruptively. So I already have arrays that are already connected to everything I need,
and I can replicate the data over and then switch the cut over so that it stays online. Yeah. I, I definitely think there's some host things in there. Sometimes, you know, depending on what that environment looks like, maybe you have to do some restarts. Obviously do Windows. But you're gonna be in complete control
of when it's executed and how it's executed. You'll just know, A, if we execute it, this is what we're going to do, and you have all the paths. And then you can take the action on top of that. Yeah. It's just simplifying.
Or maybe even you're just preceding the replication and then you're doing it manually. But it's, it's an active workload. And I, and I guess it also goes into operationally what you classify as a maintenance window and why. Because if you moved it right, you can mo- Yeah we can move it live and the there's no- Yeah connectivity loss or anything like that.
Like, your applications wouldn't know. But do you have to have a maintenance window just because you're doing that operation because that's your, your model for your, you know, for how you run your environment? Like, the data can be moved live with applications running and things. Like, no access is lost, but your organization is like, "If we do these types of operations,
we do it in a quote unquote maintenance window." Yeah. It might not be because of a technical limit. It might be just an operational- Just policy policy that you guys follow, so. And, and think about we'll help, I think, minimize the time that will take. So you can still have the window for the cut over, but maybe eliminate
all the time waiting. Yeah. It just depends on like what kind of migration are we talking about. You know? Is it a- Yeah power off, powered on? Anyways. Well, with One second. Geez. Getting a little bit of a cold. I apologize for that.
So we're at time. I just wanna wrap up quickly. Quick plug for the Everpure, community. This is our community forums website. Encourage you to go ask questions. Zane is actively monitoring on there.
I'm on there. Other customers. Yeah, feel free to join us. You can Good place for information and resources, but I just, I give this plug every time. This is not intended to replace your support. So if you are having any sort of- Yes outage- Yeah, yeah or business critical impact,
like not here. Call your support number. But for research, general information, knowledge sharing, go ahead and check that out. I just wanna say, Zane, thank you for joining.
As always, I, I always enjoy when we get a chance to do this. Yeah. And for everybody else that joined, hope you enjoyed it. I hope you got a little bit out of it. Looks like, we did get Producer Laura found a winner for the raffle, so congratulations, Chris T., on your, on your winnings there.
But I hope everybody got something out of this. Hope maybe a little educational entertainment. But until the next time- Yeah enjoy the rest- And- Enjoy the rest of your day and the rest of your week. I, I put my email in the, in the chat if you're still on, if you have any other questions or just- Yes.
Oh, wow. Glad to trade. Feel free to reach out. Yeah. That's great. Hey, I work now. With great power comes great responsibility. Yeah. Yeah. But thanks, everybody. Give me, give me grace on responding, but, happy to, happy to answer any questions.
Thanks, everybody. Until next time, like I said, enjoy the rest of your day, the rest of your week. Talk soon. Thank you, everyone.
Zane Allyn
Principal Technologist Strategist, Everpure
Jason Langer
Senior Technical Evangelist, Everpure
Array-by-array management doesn't hold up at fleet scale. Manual provisioning, one-off scripts, and per-system configs slow every request and invite misconfiguration. Everpure Fusion™ gives you a unified, intelligent control plane across edge, on-prem, and cloud, with policy-driven automation, workload placement, and built-in governance. Define presets once, and Fusion deploys every workload to the same standard, recommends where it should run, and keeps protection and compliance policies enforced across the fleet. Join us to see how teams can provision up to 90% faster and eliminate array-by-array misconfiguration.
Key Takeaways
- How an intelligent control plane turns siloed arrays into a single, policy-managed fleet across edge, on-prem, and cloud
- How presets and AI-recommended workload placement replace manual provisioning and one-off scripts
- How new workload mobility non-disruptively moves workloads across the fleet
- How compliance reporting lets you define rules once, monitor them continuously, and catch violations as they happen
We Also Recommend...
Personalize for Me
My Updates
No updates at this time.